Comparison of Diffusion of N, N-Dimethylaniline and N, N

Department of Chemistry, Georgetown University, Washington, D.C. 20057-1227 ... series expansion of the integrated form of Fick's second law for diffu...
0 downloads 0 Views 141KB Size
Macromolecules 1997, 30, 2467-2473

2467

Comparison of Diffusion of N,N-Dimethylaniline and N,N-Dioctadecylaniline in a Low-Density Polyethylene Film. Activation Energies and Detection of Two Diffusion Pathways† John A. Taraszka‡ and Richard G. Weiss* Department of Chemistry, Georgetown University, Washington, D.C. 20057-1227 Received October 28, 1996; Revised Manuscript Received January 21, 1997X

ABSTRACT: The diffusion of N,N-dimethylaniline and N,N-dioctadecylaniline from a low-density polyethylene film into 2 N hydrochloric acid was measured in real-time at various temperatures by fluorescence spectroscopy. Diffusion coefficients (D) were obtained from each data set by a best fit to a series expansion of the integrated form of Fick’s second law for diffusion through a film. The activation energies for diffusion (ED) were calculated from the D values in a temperature range above the glass transition and below the melting transition assuming Arrhenius behavior. As expected, N,N-dioctadecylaniline diffusion was slower than that of N,N-dimethylaniline at one temperature. Somewhat surprisingly, good fits of the N,N-dioctadecylaniline data sets to the series expansion require a model with two concurrent, independent diffusion processes; two diffusion coefficients at each temperature and two values of ED, 16 ( 4 and 4 ( 1 kcal/mol, are calculated. The significantly higher activation energy, similar in magnitude to the single ED of N,N-dimethylaniline, 15.7 ( 0.4 kcal/mol, is associated with the more rapid diffusional component. The data are discussed in terms of the type of sites occupied by the anilines and the influence of the two molecules on their local environments. Table 1. Some Properties of NDLDPE Films6

Introduction There are literally hundreds of “polyethylenes,” each differing from the others in degree of crystallinity, thermal history, mechanical history, number and type of chain branches, and degree and types of unsaturation.1 The noncrystalline regions consist of an amorphous part (in which chains are disordered and may be entangled and branched) and an interfacial part along the boundaries between microcrystallites and amorphous domains (in which chains are oriented in a more or less parallel fashion but do not contribute to the heat of melting2). To complicate matters further, the interfacial part may have separate regions that resemble more the crystal or the amorphous organizations.3 To understand better the dynamic nature of polyethylene, we have measured the diffusion coefficients (D) and activation energies (ED) for one or more N,Ndialkylanilines (DAA) in unstretched and stretched, native and modified films4,5 that are well-characterized.6 Among the films employed is NDLDPE, 42% crystalline in its unstretched, native state. Its other physical characteristics are included in Table 1.6 It is known that molecules like DAA are unable to enter the crystalline regions of polyethylene under the conditions of our experiments.7 There are, however, at least two families of polyethylene sites at which guest molecules can reside,8 in the amorphous and interfacial regions.4,5,7,9 If the physical processes associated with the motions leading to and from the two site types are sufficiently different, it should be possible to distinguish them kinetically by detecting two diffusion coefficients for one species in one film. To the best of our knowledge, a bimodal distribution for a diffusing guest molecule in polyethylene has not been reported previously. In our own work with lowdensity polyethylenes (i.e., e50% crystallinity), increasing the alkyl chain length in a series of DAA from † Dedicated to Waldemar Adam on the occasion of his 60th birthday. ‡ John T. Adams Undergraduate Research Fellowship recipient. X Abstract published in Advance ACS Abstracts, April 1, 1997.

S0024-9297(96)01592-6 CCC: $14.00

density: 0.918 g cm-3 thickness: 70 µm degree of crystallinity: 42% interfacial region: 3% melting temperature: 116 °C partition coefficient for DMA between film and methanol at 25 °C: 0.30 ( 0.01 CH3 groups per 1000 CH2 groups: 25 CdC groups per 1000 CH2 groups: 0.49 CdC groups per 1000 CH2 groups inaccessible to 17 wt % Br2 in CHCl3: ca. 0

methyl (DMA, molecular mass 121) to butyl (DBA, molecular mass 205) resulted in diffusion that could be described precisely by a single diffusion coefficient5a using a truncated series expansion of the integrated form of Fick’s second law.10 Here, values of D and ED are reported for the diffusion of DMA and a much larger molecule, N,N-dioctadecylaniline (DODA, molecular mass 597) in NDLDPE. Although diffusion of DMA can be described, as expected,11 by a single diffusion constant, that of DODA cannot be. A model with two distinctly different diffusion coefficients can be invoked to fit the DODA data well. Accordingly, increasing the molecular mass and size of the guest molecule permits differences between the two (principal) DAA diffusion modes predicted by the model to become detectable. The corresponding ED values suggest possible origins of each component. Experimental Section Materials. N,N-Dimethylaniline (99%, Aldrich) was distilled under vacuum and stored under nitrogen away from light. The polyethylene film from Dupont of Canada, designated NDLDPE, was 70 µm thick (Table 16). Prior to being used, film pieces were immersed in batches of chloroform for several hours to remove antioxidants and plasticizers. Methanol (Burdick-Jackson Chromopure), n-pentane (Fisher, spectro grade), hexadecane (Aldrich, 99%), and hexane and ethyl acetate (Fisher, HPLC grade) were used as received. N,N-Dioctadecylaniline was synthesized by repeated reaction of aniline with octadecanoyl chloride followed by reduction with LiAlH4. In the final step, a mixture of N-octadecyl-N© 1997 American Chemical Society

2468 Taraszka and Weiss phenyloctadecanamide (mp 55.3-59.1 m, °C; 400 mg, 0.656 mmol), 25 mL of dried THF, and LiAlH4 (74.7 mg, 1.97 mmol) were stirred and refluxed at 80 °C for 2 h under a dry atmosphere. Excess LiAlH4 was destroyed by adding ethyl acetate and distilled water. The solid was filtered under vacuum and rinsed with copious amounts of ether. The ether and THF layers were recombined and evaporated. The residue was dissolved in a minimal amount of hexane and extracted thrice with distilled water. The hexane layer was dried over anhydrous K2CO3 and evaporated under vacuum. The crude residue was purified by column chromatography (silica gel, 98/2 hexane/ethyl acetate) to yield 215 mg of product, mp 40.8-43.0 °C (lit12 mp 52-53 °C), that was 99.3% pure by HPLC analysis. FT-IR (KBr, cm-1): 2916 and 2849 (C-H stretch); 1607 (aromatic). 1 H-NMR (CDCl3 /TMS): 7.21-7.16 (aromatic, m, 2H), 6.646.61 (aromatic, m, 3H), 3.27-3.21 (N-CH2, t, 4H, J ) 7.7 Hz), 1.56 (N-CH2-CH2, broad s, 4H), 1.26 (various CH2 groups, broad s, 60H), 0.91-0.86 (CH3, t, 6H, J ) 6.6 Hz) ppm. GC/ MS: calculated M/z 597; obtained M/z 597 (1.4%); (M + 1)/z (3.9%), (M + 2)/z (1.9%). Instrumentation. UV-vis absorption spectra were recorded on a Perkin-Elmer Lambda 6 spectrophotometer in hexane or hexadecane. Spectra of films were obtained between two quartz plates. FT-IR spectra were obtained with a MIDAC FT-IR using Spectra Calc version 2.21 software. Solids were prepared as KBr pellets. A blank KBr pellet was used for background correction. Mass spectra were obtained at 70 eV on a Shimadzu QP 5000 direct inlet quadrupole mass spectrometer by Dr. Donald Weber. A Bruker 270 MHz FT-NMR spectrometer with an Apple Quadra 950 computer was used to obtain proton NMR spectra. HPLC chromatograms were recorded on a Waters 6000A solvent delivery system and a Waters Model 440 UV absorbance detector (254 nm) in conjunction with an Alltech 25 cm analytical silica gel column. A Spectra-Physics Autolab Minigrator was used to integrate the data. The eluting solvents were either 12/88 or 5/95 ethyl acetate/hexane. Emission and excitation spectra and time-dependent emission intensities were obtained on a Spex 111 Fluorolog fluorimeter with an Osram 150 W/XBO high-pressure Xe lamp and 1.25 mm slits. Samples were thermostated ((0.3 °C) using a VWR 1140 circulating constant temperature bath, and temperatures were monitored using a calibrated thermistor that was in contact with the sample cuvette. Analyses of fluorescence intensity data and fits to model equations employed the Windows versions of Quattro Pro and Origin 3.5. Procedures for Obtaining Diffusional Data.4d,13 Prior to being doped, a 1 × 2-3 cm piece of NDLDPE film was immersed in four aliquots of chloroform for 6 h each (to remove plasticizers and antioxidants) and air dried. The “clean” film was mounted taut on a glass yoke4c and immersed in a 4.7 × 10-3 M pentane solution of DODA or a 2 × 10-2 M methanol solution of DMA for 2-2.5 h (i.e., until an embibed concentration of ca. 10-3 M had been attained). In experiments with ca. 10-1 M initial DMA in a film, the methanolic doping solution contained 10-1 M DMA and the film immersion time was ca. 8 h. Initial film concentrations within a film were determined periodically by absorption spectroscopy using the Beer-Lambert law and molar extinction coefficients in hexadecane: for DMA, λmax 298 nm (max 14 730 L mol-1 cm-1); for DODA, λmax 259 nm (max 12 900 L mol-1 cm-1). The film was removed from the bath, washed with methanol to remove any surface-occluded DAA, dried under a stream of nitrogen, and rapidly placed in a thermostatted quartz cuvette containing 3 mL of 2 N hydrochloric acid (prepared from distilled water and concentrated HCl) that was inside the sample compartment of the fluorimeter. Immediately, the intensity of fluorescence was measured at a right-angle (front-face) configuration as a function of time until the rate of change in intensity was small. For DMA, λex ) 300 nm and λem ) 350 nm; for DODA, λex ) 303 nm and λem ) 335 nm.

Macromolecules, Vol. 30, No. 8, 1997

Figure 1. Emission (λex ) 303 nm) and excitation (λem ) 335 nm) spectra of 2.6 × 10-5 M DODA in hexane at room temperature.

Results Steady State Emission and Excitation Spectra. Emission and excitation spectra of N,N-dioctadecylaniline in hexane, a solvent whose polarity is similar to that of polyethylene, appear in Figure 1. Emission and excitation spectra of DMA have been reported previously.5a In 2 N HCl, the absorption (or excitation) maximum suffers a large hypsochromic shift and the intensity of emission is decreased drastically. Time-Dependent Emission Spectra. DMA and DODA were excited at wavelengths at which the NDLDPE film does not absorb. Due to the heterogeneous nature of the films, emission intensities were found to vary depending on the spot on the surface being excited. Therefore, each film was not moved during the duration of a run. However, runs in which fluorescence intensities are measured at different parts of the same piece of film led to reproducible values of D. In fact, the same piece of film can be (and usually is) employed for several months of diffusional experiments without suffering detectable changes in its properties. Emission intensity decay curves like those in Figure 2 were measured starting as close to real time ) 0 (i.e., when a doped film was placed in the aqueous acid) as possible. Practically, the experimental manipulations require a minimum of 20 s before data are recorded. Intensities at time ) 0, I0, were estimated using eq 1, a

It ) I0 - (I0 - I∞)(4/l)(D/π)1/2t1/2

(1)

modified form of the “early time” formulation of Fick’s second law,11a by extrapolating the slope of the linear region of It vs t1/2 to its y-intercept. In eq 1, It is the fluorescence intensity at time equals t and l is the thickness of the film. I∞ is taken to be the intensity when changes of